Suggestions for deploying openSSL dlls - delphi

I made a simple application that uses Indy and requires OpenSSL dlls.
I am not going to write an installer for it, so I have 2 options:
1) deploy it by copying the application exe + libeay32.dll + ssleay32.dll in the same folder
2) put libeay32.dll and ssleay32.dll in the exe resources and extract them to the applicationpath on program start (but this can may be a problem in case I am running the exe on a windowsserver 2008 or windows vista and I used the Program Files folder to "deploy")
Can you comment on those techinques or suggest a better approach?

You could stick with option 1) :D
Or, you could use a variation on number 2, but extract them to the application data path:
GetSpecialFolderPath(CSIDL_LOCAL_APPDATA))
and use
SetDLLDirectory(DirectoryPath:PWideChar)
to specify where to load dll's from
Of course, to be perfectly honest, it's quicker just to write a simple installer!
Edit:
Oh, and yes, as UnixShadow mentioned, hash the dll's so you can be sure that you're using the correct files!

Option 3: use Inno to install it! You may have a good reason for not wanting to use an installer, but in case your reason is you think it's too expensive (in time or money), you might check out Inno. I started using it earlier this year and was amazed at how easy it is to learn and use. And, it's free!
Of course the side benefit of learning Inno is that you'd have it available in the future for other projects...

Depending on the nature of the application, you could consider compiling it with Lazarus.
I know, I know, it is a far stretch, and I don't know sb who has actually done this with openssl, so it is probably too much work. Still I wanted to make a note of this theoretical solution.
This because FPC/Lazarus is mingw compatible, and can use the same (GNU linker), you could try to link mingw libs (.a's) statically.
Some people in the past have experimented with it, and I have heard people succeed in this with mysql, but unfortunately they didn't give details.
So I have no conclusive proof that it worked, which makes it a bit academical.

I would install them in the same folder (option 1), with the addition to actually MD5 the two dlls when starting my program. To try to verify that they haven't been modified. This does not help a lot when it comes to security cause a hacker might as well modify my MD5 sum I would compare against. But it would at least stop the program from using corrupt dlls. Ofcourse this would also block out the possibility to update the dlls without updating the main application (on good an bad).

Related

Detailed Explanation on "Re-Hosting" and "Retargeting" both for compilers and binary data (such as .exe or .obj)

Sometimes in s/w companies, customers provide data in multiple formats. There are linkable and executable data that are said to be "Rehosted" and compiled object files that are said to be "Retargeted". I am trying to understand what rehosting and retargeting mean in this area. Is it similar to the Bootstrap theory in computer science? I have the understanding of the following process (if not incorrect):
PROBLEM:
I need to write a compiler for a new language called "MyLang" to run on PowerPC
Solution:
1. I need to write a compiler for a language "MyLang-Mini"; a subset of "MyLang" to run on PowerPC.
2. I need to write a compiler for "MyLang" using "MyLang-Mini" to run on PowerPC.
3. I run the compiler obtained from no. 1 through the compiler obtained from no. 2 to
obtain the compiler for MyLang to run on PowerPC.
IN BESPOKE "T" DIAGRAM (...ISH):
MyLang PowerPC MyLang PowerPC
MyLangMini MyLangMini PowerPC PowerPC(instr.)
PowerPC(instr.)
What I am getting confused about is rehosting and retargeting. How are they coonected to this concept? What am I rehosting and retargeting if I have some binary data such as .exe or .obj? I would appreciate some detailed explanation if possible please!
I know that this will embark onto "CROSS-COMPILERS", but would prefer expert opinions to be sure.
Thanks in advance.
I now know that in s/w engineering:
REHOSTING - If you have a third-party application linkable/executable that requires usage on your host machine, you do rehosting. The target in this case are most often the same (OS platform, processor, etc.). In worst case, there is a virtualisation required. The rehosted application will run as if it was one of the application running in the host machine
RETARGETTING - If you have a third-party source code, you might need to recompile that to match with your target environment. It may also be that you have third-party .o or .obj compiled models and you want to link them with your source code (retargeted) in order to host it on a host machine. Just like REHOSTED application, it will be as if the application was installed on the host machine.
It will be good to know how this is similar to the compiler rehosting and retargeting. Sorry, I am a newbee is this area and will appreciate even a slap on the wrist.

Building VB6 projects without registering

We have an old VB6 project that uses ActiveX controls, some of which we build and others we get from third-party vendors.
Currently, we use a .csproj project which does the following,
Execute regsvr32 to register the OCXs
Execute vb6 to build the VB6 project
Execute regsvr32 to unregister the OCXs
This registering/unregistering is ugly and is a bit of a pain for local developer builds with UAC enabled. Is it at all possible to build a VB6 project without having to register any controls?
I apologize if this has already been asked before. The only similar questions I was able to find were about how to build VB6 projects, and answers to these mention the same solution of register, build, unregister.
It sounds like these people are merely working on clients of these OCXs rather than modifying and recompiling the OCXs themselves.
If so, you should be administering the installation of these libraries just as you administer the VB6 development system itself. This means each workstation needs to have the control suites you are using installed once (well, and maintained when new releases are placed into use). Installers for developer libraries deploy things like .DEP files as well as design-time license key registry entries, so using regsvr32 shouldn't be considered a viable strategy anyway.
If you set the developer workstations up properly and maintain them there isn't any reason to be registering and unregistering such things.
It means the original developers probably did not set the "binary compatibility" correctly. Which means the VB6 dll's get a "new com guid" every time they are built.
Which means your original VB6 developers were probably a bunch of hacks.
You can read the section here on Binary Compatibility.
http://support.microsoft.com/kb/161137
Get in a time machine and go back and punch the person in the face who said "We don't need
to work out the binary compatibility issues now, we'll just unregister and re-register the components... Easy Peezey!"................
If I'm wrong, please let me know. But every time I've seen "unregister the com" and "re-register the com".........it goes back to that brainiac decision.
Here is a longer discussion on it:
http://www.techrepublic.com/article/demystifying-version-compatibility-settings-in-visual-basic/5030274
EDIT:
If the ocx's are not changing........then you should only have to register them once on the build machine once.
The direct answer is no, it is not possible to compile a VB6 project with OCX dependencies without those dependencies being registered.
Furthermore, the act of compilation itself involves VB6 attempting to register what it has just built (unless you are compiling to an EXE). This generally requires the VB6 IDE and/or its compiler to run with "admin" permissions. Therefore the permissions are a hard to avoid issue regardless.
I believe these issues can be obfuscated by the fact that VB6 itself (the IDE and/or the runtime) will sometimes try to automatically register certain things for you, but will keep silent when it does so.
You should probably create a different process to setup a development PC from the build process you use from deployment. This may "feel" wrong especially if you have experience with other programming environments, but I would stress that VB6 can be very painful & problematic to work with and so pragmatism is generally in order.
On the development PCs: Setup all the unchanging dependencies once (and document them) and then leave them alone (as noted in another answer.) When weird dependency problems occur, verify the PC is setup correctly before doing anything else.
If you have all the sources to your dependencies, then I would consider if you can actually run them all in a VB6 project group (VBG) and not compile them at all. (A VBG is akin to a .NET solution though far less powerful.) I do this often and it cuts out a lot of wasted time. Developers don't necessarily need code compiled to EXE / DLL / OCX - they often just need to be able to run it in the IDE.
On the build PC: If you can always start with a clean environment, like in a virtual machine, then I think its actually a good idea to register everything from scratch in an automated fashion as this helps to verify nothing is missing or mismatched. Re-using the same build environment without doing this can mask problems when some dependency has changed in source control but still exists on the build machine. On a VM generally permissions aren't a limiting factor.
Notes:
If you are building an EXE, VB6 does not require any elevated permissions, as far as I can recall.
Running code in the VB6 IDE does not either.
[Caveat 1]:
It may technically be possible to create a side-by-side application manifest file for VB6.exe itself and include in that manifest whatever dependencies you need, thereby avoiding having to register them.
But this would fall well outside of the normal ways to use VB6 tools - its a hack - and possibly is not worth the potentially large effort. I don't think I've ever seen a working example and so I don't recommend this as a practical solution, but mention it for completeness.
Maybe in some locked-down corporate IT scenario this could pay off... maybe. In that scenario doing dev work in a VM might be a better option though.

How to migrate Delphi or clone Delphi registry settings?

I have two PCs both with XE2. I thought that I had installed identically on both but have problems installing 3rd party packages on one while the other is just fine.
I want the same on both anyway. The easist would probably just to "migrate" the working set-up by moving in into my Dropbox folder. Can I do that? If so, how?
If not, can I (easilly) backup my registry settings on one machine and then import them on the other?
I suppose I could just sort out the problem on the one PC, but am not having much luck so far. I would rather invest the time in only having one Delphi setup. And since I am moving lots of other stuff to DropBox anyway ...
The tool for this is now built into Delphi XE8 and higher.
It's found here:
C:\Program Files (x86)\Embarcadero\Studio\20.0\bin\migrationtool.exe
Online documentation:
http://docwiki.embarcadero.com/RADStudio/Rio/en/Settings_Migration_Tool
Install CnPack wizards from http://www.cnpack.org
From the CnPack toolbar select IDE Config Backup/Restore (image below) and save this file somewhere safe
Copy the components to the second delphi machine . Keep the exact same directory structure.
I store my components as follows this helps backing up, moving etc., but you can use your own structure
D:\components_bds\DCU
D:\components_bds\BPL
D:\components_bds\ComponentsThemselves
Use the restore config file from CnPack to restore your components on the new machine
This is also useful if your testing components that you plan to remove later and keeping a backup of your installation incase something goes wrong you can save time with new delphi installation if hard drive dies. Keep a copy on flashdrive or somewhere safe
You may compare/diff the config file created by cnPack using a tool like Beyond Compare and see what the differences are to find out why third party components give problems on one of the machines. It may be a Delphi registry/installation problem or a problem in the paths of the thirdparty components. Components need to be installed in an order perhaps it did not find the needed dcu or dll it depends on.
I don't know of any way to do so with DropBox. Here's an old post I made (related to Delphi 7, but with correction of registry keys still applicable) in the CodeGear newsgroups; hopefully it will help.
(It probably goes without saying, but back up the existing registry settings on the destination machine before starting by using RegEdit and exporting them, just in case. You'll at least be able to get back to the point you're at now if something goes wrong by deleting the imported entries and then importing the saved ones.)
You can't, without some difficulty anyway. (Especially if you have
third party components installed, as they may have placed files in the
%SYSTEM% folder you may not know about.)
You may be able to (for going from the old computer to the new
computer running the same exact version of Windows!) by exporting the
registry keys under HKCU\Software\Embarcadero and
HKLM\Software\Embarcadero from the old machine, and then after
installing Delphi on the new machine (in the exact same folder
location) importing that registry file.
Many of the compiler, linker, and other settings are configured on a
per-project basis, and should transfer over when you move your source
code to the new machine.
Third-party components are a problem, as I mentioned above. You may be
able to get away with using the registry export/import if you copy
each third-party component set from the old computer into exactly
the same location on the new machine before importing the registry
file. You'll probably have to track down some .BPL files that end up
in the $(BDS)\Bin and possibly other folders under the $(BDS)
tree; the IDE will tell you about missing stuff when you try and start
it. Make sure you answer "Yes " when asked if you want to try and load
it again next time!
Most of my development is hobby stuff or wannabe releases. Instead of dying trying to move my XE2 Pro from my Dell Inspiron N7110 Win 7 machine to my new Win 10 SSD machine, I'm seriously thinking of switching to Lazarus. I've used Lazarus 2.x with Indy 10, ZeosLib, and Firebird and successfully created a working distributed internet system. I also created Lazarus version of my XE2 Blackjack program. When compared to XE2, Lazarus (IMO) has only two weakness and neither are deal breakers for me. BTW, I have successfully duplicated Lazarus (with all installed components) from one machine to another simply by copying and pasting the Lazarus directory and it works. Try that with Delphi.
Sam

Single-user source control?

Searching with '[Delphi] "source control"' didn't return much, so here goes: For those of you Delphi programmers working on your own, what source control do you like? I know about TortoiseSVN (which can be used without a server), but I'd like to know if there are better options before choosing it.
Thank you.
SVN will be more than enough as you'll mostly use it for backing up and diff'ing versions.
You can use VisualSVN Server and TortoiseSVN with Delphi IDE Integration. All of these are freeware. If you wish to use any other freeware SCM with Delphi IDE integration you'll need to buy SourceConneXion or Athlant. Also, if you're using D2009 you'll need to convert strings to ansistrings in order to get TortoiseSVN addon for delphi to work.
I use Mercurial, and have used Bazaar and Perforce (free for small teams). All are good, but nowadays I tend to prefer Mercurial. Mercurial comes with its own tortoise
which, while not quite as polished as TortoiseSVN, is perfectly usable:
(source: sourceforge.net)
For users of RAD Studio XE or XE2, Uwe Schuster has published an IDE add-on called Version Insight Plus that adds Mercurial and Git support to the IDE's built in version control support.
For all the systems I have mentioned, no central server is required.
I would recommend Git which is free & open source and:
Doesn't require you to even set a central server even if you want add more developers or machines.
Is Extremely Fast (imo)
Encourages the use of branches
I use it for almost every new project, even when it's just me on the project. It's an extremely fast distributed version control system and was written by Linus Torvalds and is now used in high profile projects like the Linux Kernel and Ruby on Rails.
Git isn't hard to use from the Command line but also has it's own "Tortoise" package (TortoiseGit - albeit not as polished as it's SVN cousin).
SourceGear Vault is free for a single user. I like to use the external client because I work in different environments, but If you want Delphi IDE integration, you can use Source ConneXion (not free).
You might look at some of the responses here (Stack Overflow) if you do have Delphi in mind. As mentioned in the answers there, I use Team Coherence which integrates very nicely with Delphi's IDE, and is aware of Delphi file groups (.pas with .dfm etc). I think it's written in Delphi too.
Having said that, the other comments already made are true - you shouldn't really let your choice of language dictate your choice of VCS.
Even though you've said single user, I would look for a solution that allows you to easily host it on a server/other desktop machine, so you have separated your development machine from your source repository. I'm probably teaching granny to suck eggs but you ought to check that whatever you choose can be easily backed-up too (even to a USB key or external drive would be fine). :-)
I know you have asked for Source Control, but if you are always planning to be a 'single developer' you might like to consider an automatic backup solution like AJC Active Backup instead. Yes, you lose the ability to check in and out specific versions, but at the same time you avoid the need to check stuff in and out all the time when it is only you working on a project. And you can recover or diff any previous source files by date for as far back as you care to configure. It's very much a set-and-forget solution, until you need to recover something, when it is invaluable. And the archive itself can be backed up in the normal way.
I use TortoiseSVN, but store my repositories as files on my laptop (on a share that can be accessed by mutilple VMs), not managed by a server. This means that I have full access to the repository regardless of server access, but also allows me to easily backup the repositories to the company file server when connected.
Since I am the only programmer I have not needed a separate server so far. I have seen Nick Hodges (Delphi Development Manager) say very nice things about VisualSVN Server, and I aim to check this out, to see how easy it is to backup/restore the repository: if that can be automated, and done very quickly, then I will probably adopt that.
Longer term I am going to look at the various flavours of distributed VCS, as that may be better suited to multiple laptop-based developers; I'm not sure yet, as I've not reached that page of my to-do list ;-)
One of the things supposedly in the pipeline for Delphi is integrated support for VCS. I've no details on that, or on any implications for current users of the various traditional or distributed VCS.
The language you use doesn't really matter in the choice of the SCM you will use.
It can matter if your favorite IDE supports or not this SCM.
TortoiseSVN is just a svn client, if you choose svn you can have several client.
I use Tortoise, and subclipse for committing in the same svn repository.
I would suggest using SVN server on a separate machine (either VisualSVN as suggested before or CollabNET Subversion Server) and TortoiseSVN with JVCL integration expert (also as suggested before).
Besides getting all the good stuff from the version control, you'll also automatically have backup on a different computer, which is always a good thing.
I know this will get down voted, but I feel it has to be said.
I've used Version Control software for many years at my job, and it is required when multiple people work together - to make absolutely sure that no two people overwrite one-another's work.
But for my personal development at home I don't use one. I find them overbearing and inconvenient for a simple one-person project. And I've tried several packages including Source Gear Vault and some others that other people have mentioned.
What I do instead is at every significant change to my program that I might want to go back to, I make a copy of the entire program directory. If I screw up in the next change, I can go back to the copied directory.
I also permanently keep copies of all my directories of every release of my program. If I ever need to do comparisons between my current version and previous releases, or between two different previous releases, I use Beyond Compare by Scooter Software - a simply great tool for diff-ing and copying changes between versions. If you use Source Control software, Beyond Compare will integrate into it nicely.
So I mention this simply because I have often heard everyone pushing Version Control Software, even for the individual developer. For some of us, it's overkill.
You may want to consider this simpler solution.
I use Bazaar with Delphi and it works well, especially for solo developer workflows. It has it's own TortoiseBzr but it isn't as good as some of its cousins so I recommnend the excellent Bazaar Explorer instead. One of the great things about Bazaar is it's flexibility and the ease with which you can change workflow or include another developer in a project that started off solo.
A lot of articles about Bazaar will tell you it is a lot slower than Git or Mercurial. They are out of date, now it takes a similar time to acomplish operations as they do and is faster on some operations.
My team use StarTeam and I have never used something else like SVN so it is hard for me to do a good compare. Starteam has both positive and negative sides.
Pros:
You can register Change requests and
connect them to checkins. This make
changes more trackable.
I feel more comfortable with a real GUI compared to commandline or some Shell
extension.
Cons:
Expensive, as most products from
Borland...
Latest version use Java. It is almost like the previous Win32 version and they have added more features, but I still feel that it is some slower and more memory hungry than before.
GUI could be a bit more intuitive.
I'm not big on version control, but use SVN/Tortoise and am quite happy with it. The main benefits that I see for a single developer is being able check out older versions of the application, and use multiple computers for development (desktop / laptop) - other than the obvious backup benefit. I tend to not need it for diffing files, as I find the version control tools within the Delphi IDE fairly good for this - provided the "bug" was introduced recently.
I think most Delphi developers use SVN/Tortoise (making it a fairly safe choice) - but there are trendyer options such as Mercurial and Git.
As a single developer, I have used Perforce for a number of years. It has been great. Aside from a CLI, you can use the P4V client. There's also an explorer plugin available, as well as Delphi IDE integration. I actually just found this free integration on Torry yesterday:
p4delphi
I've installed it in Delphi 2010, and it works pretty good.
Perforce is free for two user, it would require you to install perforce server, but the server footprint is very, very small. you can use P4V (Perforce visual client) which is far more convenient than tortoise, it also provide Windows explorer extension to be have just as tortoise (but you could skip this installation). Delphi integration could be done using P4Delphi.
I am using subversion, bug tracking and simple to use project planning for my latest delphi project from an web service provider. http://www.unfuddle.com has a free account for projects less than 200mb with 1-2 developers and the user interface is easy to use.

Have COBOL compiled Program -- Missing Source

We have a very important application, but so far no source code. The application was written in COBOL and a compiled version is in our production system and is running.
However, we need to migrate to a new server, and new cobol compiler. We're under the impression that we need to recompile the code to get it to work on the new server. Running the exising compiled program gave runtime memory errors.
We have some source code for the program, but it is old. Not sure what the diff is between it and the compiled program.
Okay, so the question -- what should we do?
Time is not on our side, since we have to send our old server back to get credit for it. Ideas, suggestions, crazy or otherwise? (source control is obvious and its not up to me to do it, so save the lectures)
Create an image of your old server. Then run the old server as a virtual machine on your new server.
However, I agree a better option is probably to keep your production server.
In the short run it would probably be cheaper to arrange to keep the old server. In the semi-long run, you need to make time and budget to reengineer the program, either re-write it or see how much effort it would be to hack the old code into shape doing what the program currently does.
It's sadly. You should consult the Source Recovery Company
If your source code is relatively close to the compiled version, try this:
decompile new version into assembler
compile the old source code into assembler
compare
reconcile as best you can the differences from new version with old version, into the old souce code
repeat
To augment this, and probably as a second step, as it will bring the source code farther from the new compiled version, test with input data and just try reverse-engineering based on the output what would be needed to create that output. The more test input data you have the better this could work.
Good luck!
(I'm not a COBOL programmer but..)
If you know what version the compiler was that compiled the original program, you could at least compile the old cobol source; if the compiled versions is identical you know the source actually is the current version.
If they differ, you could try to (somehow) decompile, or at least disassemble, the working compiled version and the freshly compiled version and use a diff tool to get an idea of how big difference there is.
crazy sugestion: COBOL DECOMPILER --> SOURCE --> NEW COBOL COMPILER...?
(edit: http://juggersoft.com - PAID cobol decompiler)
if you have the .int (intermediate) binary files you can just run on the new server, if not, them you musto to recompile.
The program could have been produced by an external resource and that person or software house or organisation could have the latest source in their repository. It may be held by your parent organisation if you have recently merged, or may be in a different or backup computer installation in your organisation. There may be a copy on the developer's user account and may not have been sent to the production or live site or someone from head office has a copy to assess and try to resolve the situation. You may have success if phone those people or you could always talk to the installation computer operator or support staff and see if they have one on mag tape, CDROM or other backup storage.

Resources