[gentoo-dev] suspicious code in gnustep eclasses

[gentoo-dev] suspicious code in gnustep eclasses

["Paweł Hajdan, Jr."]

[-- Attachment #1: Type: text/plain, Size: 1329 bytes --]

This is from gnustep-base.eclass:

> egnustep_doc() {
>     if [[ -d ./Documentation ]] ; then
>         # Check documentation presence
>         cd "${S}"/Documentation
>         if [[ -f ./[mM]akefile || -f ./GNUmakefile ]] ; then
>             emake "${GS_ENV[@]}" all || die "doc make failed"
>             emake "${GS_ENV[@]}" install || die "doc install failed"
>         fi
>         cd ..
>     fi
> }

Shouldn't those cd calls above rather be pushd/popd? It seems the above
assumes that CWD is "${S}" when egnustep_doc is executed, which is
probably true, but pushd/popd seems just safer.

Also, instead of ./Documentation, "${S}/Documentation" could be used.

This is from gnustep-2.eclass:

> RDEPEND="${DEPEND}
>     debug? ( >=sys-devel/gdb-6.0 )"

Is there some gnustep crash-reporting tool that uses gdb? I think it's
reasonable for USE="debug" to influence how things are compiled, but
unless gdb is required for something to work, it should be up to the
user to install or not install gdb.

In case something is broken with <gdb-6.0, please consider two points:

- there is no <gdb-6.0 in the tree now
- you could add a blocker on <gdb-6.0 instead, which is not going to
disrupt developers because there is no such version in the tree anyway,
and we have up-to-date systems


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 203 bytes --]

Re: [gentoo-dev] suspicious code in gnustep eclasses

[Fabian Groffen]

[-- Attachment #1: Type: text/plain, Size: 1692 bytes --]

On 30-03-2012 13:00:33 +0200, "Paweł Hajdan, Jr." wrote:
> This is from gnustep-base.eclass:
> 
> > egnustep_doc() {
> >     if [[ -d ./Documentation ]] ; then
> >         # Check documentation presence
> >         cd "${S}"/Documentation
> >         if [[ -f ./[mM]akefile || -f ./GNUmakefile ]] ; then
> >             emake "${GS_ENV[@]}" all || die "doc make failed"
> >             emake "${GS_ENV[@]}" install || die "doc install failed"
> >         fi
> >         cd ..
> >     fi
> > }
> 
> Shouldn't those cd calls above rather be pushd/popd? It seems the above
> assumes that CWD is "${S}" when egnustep_doc is executed, which is
> probably true, but pushd/popd seems just safer.

Go ahead.

> Also, instead of ./Documentation, "${S}/Documentation" could be used.

Given the following cd, I tend to agree.

> This is from gnustep-2.eclass:
> 
> > RDEPEND="${DEPEND}
> >     debug? ( >=sys-devel/gdb-6.0 )"
> 
> Is there some gnustep crash-reporting tool that uses gdb? I think it's
> reasonable for USE="debug" to influence how things are compiled, but
> unless gdb is required for something to work, it should be up to the
> user to install or not install gdb.
> 
> In case something is broken with <gdb-6.0, please consider two points:
> 
> - there is no <gdb-6.0 in the tree now
> - you could add a blocker on <gdb-6.0 instead, which is not going to
> disrupt developers because there is no such version in the tree anyway,
> and we have up-to-date systems

I think the version is because GNUstep is written in Objective-C.  That
said, I think your blocker approach would be fine.


-- 
Fabian Groffen
Gentoo on a different level

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: [gentoo-dev] suspicious code in gnustep eclasses

["Paweł Hajdan, Jr."]

[-- Attachment #1: Type: text/plain, Size: 881 bytes --]

On 3/31/12 8:45 AM, Fabian Groffen wrote:
> On 30-03-2012 13:00:33 +0200, "Paweł Hajdan, Jr." wrote:
>> This is from gnustep-base.eclass:
>>
>>> egnustep_doc() {
>>>     if [[ -d ./Documentation ]] ; then
>>>         # Check documentation presence
>>>         cd "${S}"/Documentation
>>>         if [[ -f ./[mM]akefile || -f ./GNUmakefile ]] ; then
>>>             emake "${GS_ENV[@]}" all || die "doc make failed"
>>>             emake "${GS_ENV[@]}" install || die "doc install failed"
>>>         fi
>>>         cd ..
>>>     fi
>>> }
>>
>> Shouldn't those cd calls above rather be pushd/popd? It seems the above
>> assumes that CWD is "${S}" when egnustep_doc is executed, which is
>> probably true, but pushd/popd seems just safer.
> 
> Go ahead.

I've checked in this and all other changes to CVS, after testing gnustep
packages locally.

Paweł


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 203 bytes --]