Smalltalk › Squeak › Squeak VM

[OpenSmalltalk/opensmalltalk-vm] Check sanity of curled gzip data (#326)

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

8 messages Options

David T Lewis

[OpenSmalltalk/opensmalltalk-vm] Check sanity of curled gzip data (#326)

The getGoodSpur*VM.sh scripts assume that the VM mentioned in the
topmost notification, is always there on bintray. This is not
always the case. Unfortunately, bintray will not return a 404 code
on a 404 error, but instead give a 200 with garbage data. While we
can't do anything about that, at least make an attempt to bail if
the file doesn't even look like gzip-compressed data.

You can view, comment on, or merge this pull request online at:

https://github.com/OpenSmalltalk/opensmalltalk-vm/pull/326

Commit Summary

Check sanity of curled gzip data

File Changes

M image/getGoodSpur64VM.sh (9)
M image/getGoodSpurVM.sh (8)

Patch Links:

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

David T Lewis

Re: [OpenSmalltalk/opensmalltalk-vm] Check sanity of curled gzip data (#326)

This looks good, however shouldn't it "exit 1" on error so the calling script knows not to proceed?

Bonus points :-): It would be really nice if it could retrieve the previous version (but I haven't looked at bintray to see if they is practical).

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

David T Lewis

Re: [OpenSmalltalk/opensmalltalk-vm] Check sanity of curled gzip data (#326)

In reply to this post by David T Lewis

@gcotelli commented on this pull request.

In image/getGoodSpurVM.sh:

> @@ -52,6 +52,14 @@ else
 				exit
 			fi
 			curl -L "$URL" -o "$LATESTVM"
+			# bug in bintray: on a 404 error,
+			# it will return a 200 with garbage data
+			if [[ $(file "$LATESTVM" | grep 'gzip compressed data') ]]; then 
+			  tar xzf "$LATESTVM"
+			else
+			  echo No gzip data at "$URL"
+			  exit
+			fi
 			tar xzf "$LATESTVM"

This line seems to be duplicated. It's already contained in the if

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

David T Lewis

Re: [OpenSmalltalk/opensmalltalk-vm] Check sanity of curled gzip data (#326)

In reply to this post by David T Lewis

@shingarov pushed 2 commits.

f081826 Merge branch 'Cog' of github.com:shingarov/opensmalltalk-vm into Cog
1015842 Merge branch 'Cog' of github.com:shingarov/opensmalltalk-vm into Cog

—
You are receiving this because you are subscribed to this thread.
View it on GitHub or mute the thread.

David T Lewis

Re: [OpenSmalltalk/opensmalltalk-vm] Check sanity of curled gzip data (#326)

In reply to this post by David T Lewis

Looks good. I'll leave it a bit longer in case there is any more feedback and then merge it in.

Thanks,
Alistair

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

David T Lewis

Re: [OpenSmalltalk/opensmalltalk-vm] Check sanity of curled gzip data (#326)

In reply to this post by David T Lewis

Merged #326 into Cog.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

David T Lewis

Re: [OpenSmalltalk/opensmalltalk-vm] Check sanity of curled gzip data (#326)

In reply to this post by David T Lewis

Hi Boris, I’m wondering if the right thing to do is to build a vm that includes the processor simulator plugins in the relevant ../build.* directories. The makefiles are set up to build the processor simulator plugins only if the support libraries (the processor simulator code itself) have been built, so by default the CI does not build them.

I would prefer that we set up a CI to build the processor simulator plugins and download and install these as part of the getGoodSpurVM scripts, but that’s work I haven’t found time for and at least on Mac installing a plugin in the .app will break the signature and provoke a warning on launch. So building a vm in place, while slow and not yet automated might be the right approach. It does mean running the conf.COG scripts and make in the relevant build./{bochsx86,bochsx64,gdbarm32} directories, none of which is as smooth as it might be. What do you think?

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

fniephaus

Re: [OpenSmalltalk/opensmalltalk-vm] Check sanity of curled gzip data (#326)

On Sun, 23 Dec 2018 at 7:44 pm, Eliot Miranda <[hidden email]> wrote:

Hi Boris, I’m wondering if the right thing to do is to build a vm that includes the processor simulator plugins in the relevant ../build.* directories. The makefiles are set up to build the processor simulator plugins only if the support libraries (the processor simulator code itself) have been built, so by default the CI does not build them.

I would prefer that we set up a CI to build the processor simulator plugins and download and install these as part of the getGoodSpurVM scripts, but that’s work I haven’t found time for and at least on Mac installing a plugin in the .app will break the signature and provoke a warning on launch.

These kind of plugins (and probably others as well, say SqueakSSL) can easily be pulled/downloaded into the plugins directory from sub-repositories as part of the main CI builds. This way, they will also be signed correctly.

Fabio

So building a vm in place, while slow and not yet automated might be the right approach. It does mean running the conf.COG scripts and make in the relevant build./{bochsx86,bochsx64,gdbarm32} directories, none of which is as smooth as it might be. What do you think?

—

You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.