[Rubyinstaller-devel] Yet another Windows installer

Luis Lavena luislavena at gmail.com
Sat Mar 22 11:16:59 EDT 2008

On Sat, Mar 22, 2008 at 6:33 AM, Michal Suchanek <hramrach at centrum.cz> wrote:
>  I am not arguing against Rake, I am just saying that there are other
>  tools that could do the job, and thus there are people who aren't
>  familiar with Rake.

Don't take my comments personal. I self learned english and sometimes
what I express in words are not actually 100% accurate on what I think
it should express :-)

I had used in the past a series of batch files to perform this, and if
you take a closer look of the current code in installer2 in the
rubyforge repository, you will see Curt and Andy already used
something similar to bootstrap building One Click Installer

By "people" I meant other Ruby developers, working on Windows, willing
to collaborate in the project. The end-user of the project wouldn't
care about all the effort of download MinGW or build everything from
scratch, they will only use the generated output (the MSI installer).

>  >
>  >  If you have used the garbagagecollect builds for your "Yet another
>  >  Ruby Installer for Windows", then you know there is no documentation
>  >  or way to recreate the process locally, and thus there is no way to
>  >  simple fix the problems the current Ruby for Windows faces.
>  >
>  >  Then you're stuck each time to the releases made by somebody else and
>  >  reply lots of tickets and support request when things don't work out
>  >  of the box, like WIN32OLE in 1.8.6-p110 (which was replaced soon
>  >  enough by p111)... I think you get the picture.
>  Yes, I have seen the earlier discussions about installers. It's
>  actually not mine "yet another installer", I was just replying to the
>  thread. And I am aware of the grief with binary-only releases :-/

Sorry about that, got lost in the translation :-P

The dependency on external and non directly related developers to
release fixes is a no-no. As example, take Mongrel project. I shared
the most as possible of my Windows environment setup with the other
Mongrel developers, so they don't need me to be around when a
bug/security fix show up.

Right now, we depend 110% on garbagecollect releases, but the few
times I politely asked the developers for his scripts (to recreate the
build environment locally) he didn't replied back.

So I ended creating, from scratch again, the build scripts for VC6 and
VC8 at that time, just to found it too complicated and decided to go
for a Pure Ruby alternative :-)

>  >
>  >  1) after you unpack latest.zip (I actually name the folder
>  >  installer3.dev, anyway) you jump in and perform the 'rake download'
>  >  task.
>  Well, the zip archives contain an installer3 folder ...

Oh yes, the zip files contains installer3 folder, I meant the Bazaar branch :-)

>  7) and then I download a new latest.zip, and since I am not sure
>  nothing was removed between the different version the only safe way is
>  to start with a fresh copy of the installer3 folder.

That will end download *everything* again... please don't do that... a
huge hit to SF server is not polite :-(

>  BTW stuff breaks if the path to installer3 contains a space, this
>  could be mentioned in README it it is not already.

Hehehe, I just get used to not use spaces in my path that forgot about it.

I'll rewrite README for better reading of requirements and limitations.

>  >
>  >  I'll fix the extract utils to expand the downloaded zlib package.
>  Yes, that's what I had in mind. There's even a separate download for
>  zlib1.dll only.

We already have it, the download is zlib123-dll.zip

I'll patch the extract utils in the following minutes.

>  I guess it is time to add it because with rubygems and rake added you
>  have enough to rebuild ruby with the image that is built by the
>  recipes.

You're correct, maybe I'll do that before putting all the pending
dependencies in place :-P

>  Also you could release the image as .zip and somebody would perhaps
>  try it out and test their favourite gems.

I'm still concerned by readline test being broken and that this
generates IRB uses 50% of the processor being idle!
(yes, that is a huge bug that need to be backported to 1.8 from trunk
branch of ruby).

>  I guess iconv and openssl are important but curses are broken anyway.
>  Actually they might work on some language versions of Windows that do
>  not (and cannot) use multibyte characters but then you have the
>  additional problem of finding out the encoding that's currently in
>  use.

We need to determine what built-in extensions are really used. OpenSSL
is a must since signed gems (like Mongrel) will not work without it.

for Iconv I think will follow the suggestions from _why and use the
win32 alternative, the gnuwin32 packages for these tools are older and
unmaintained :-(

Thank you for your feedback Michal!
I really appreciate it and it give me a push to fix these things! :-)

Regards and have a nice weekend!
Luis Lavena
Multimedia systems
Human beings, who are almost unique in having the ability to learn from
the experience of others, are also remarkable for their apparent
disinclination to do so.
Douglas Adams

More information about the Rubyinstaller-devel mailing list