<div dir="ltr"><div><div>In that case I agree there's no urgent need for us to put out an 8.4.3. But we should explain the rationale carefully, it looks like this is a sensitive topic.<br><br></div>Cheers<br></div>Simon<br></div><div class="gmail_extra"><br><div class="gmail_quote">On 13 May 2018 at 17:00, Gershom B <span dir="ltr"><<a href="mailto:gershomb@gmail.com" target="_blank">gershomb@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div style="word-wrap:break-word;line-break:after-white-space"><div id="m_-2728895057694851186bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px;color:rgba(0,0,0,1.0);margin:0px;line-height:auto">Update: as per <a href="https://github.com/commercialhaskell/stack/issues/4019" target="_blank">https://github.com/<wbr>commercialhaskell/stack/<wbr>issues/4019</a> it turns out that stripping is in fact enabled by default in stack, so this bug only hits when people _turn off_ stripping in their config files (which in stack just happens to also be coupled with turning on -g).</div><div id="m_-2728895057694851186bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px;color:rgba(0,0,0,1.0);margin:0px;line-height:auto"><br></div><div id="m_-2728895057694851186bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px;color:rgba(0,0,0,1.0);margin:0px;line-height:auto">So a fresh installed stack works on ubuntu, and only breaks when people change stripping settings. Anyway, this seems to lend more weight to the “no release necessary” side of things.</div><span class="HOEnZb"><font color="#888888"><div id="m_-2728895057694851186bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px;color:rgba(0,0,0,1.0);margin:0px;line-height:auto"><br></div><div id="m_-2728895057694851186bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px;color:rgba(0,0,0,1.0);margin:0px;line-height:auto">—Gershom</div></font></span><div><div class="h5"> <br> <div id="m_-2728895057694851186bloop_sign_1526226950123761152" class="m_-2728895057694851186bloop_sign"></div> <br><p class="m_-2728895057694851186airmail_on">On May 13, 2018 at 4:11:06 AM, Gershom B (<a href="mailto:gershomb@gmail.com" target="_blank">gershomb@gmail.com</a>) wrote:</p> <blockquote type="cite" class="m_-2728895057694851186clean_bq"><span><div style="word-wrap:break-word;line-break:after-white-space"><div></div><div>








<div id="m_-2728895057694851186bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px;color:rgba(0,0,0,1.0);margin:0px;line-height:auto">
On May 13, 2018 at 3:59:04 AM, Simon Marlow (<a href="mailto:marlowsd@gmail.com" target="_blank">marlowsd@gmail.com</a>) wrote:</div>
<div>
<blockquote type="cite" class="m_-2728895057694851186clean_bq" style="font-family:Helvetica,Arial;font-size:13px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none">
<div>
<div>
<div dir="ltr">
<div>
<div><span>Will Stack also be broken for other GHC versions on
Ubuntu 18.04, or does it only enable -g for 8.4+? i.e. even if we
roll a new 8.4.3, will people still using 8.2.x have
problems?</span></div>
</div>
</div>
</div>
</div>
</blockquote>
</div>
<p>Indeed, as best I undesrtand it, it will be broken on the 8.2
series as well. So anyone pinned to any older ghc (via a particular
lts or daily) will continue to have this problem. From what I can
tell, the -g flag is added basically universally.
C.f. <a href="https://github.com/commercialhaskell/stack/blob/master/src/Stack/Build/Source.hs#L147" target="_blank">https://github.com/<wbr>commercialhaskell/stack/blob/<wbr>master/src/Stack/Build/Source.<wbr>hs#L147</a></p>
<div>
<div>
<blockquote type="cite" class="m_-2728895057694851186clean_bq" style="font-family:Helvetica,Arial;font-size:13px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none">
<div>
<div>
<div dir="ltr">
<div><span><br class="m_-2728895057694851186Apple-interchange-newline">
The "purist" side of me agrees with you that this is a downstream
issue.  However, Stack doesn't have its own builds of GHC
unlike Ubuntu, so it can't easily fix this by patching GHC. They
could fix Stack to not use -g on affected distros / GHC versions
though.</span></div>
</div>
</div>
</div>
</blockquote>
</div>
<p>As far as I know, stack does
produce its own builds these days, although they should be
conferred with, to confirm. In particular, see <a href="https://raw.githubusercontent.com/fpco/stackage-content/master/stack/stack-setup-2.yaml" target="_blank">https://raw.<wbr>githubusercontent.com/fpco/<wbr>stackage-content/master/stack/<wbr>stack-setup-2.yaml</a> which
is where it draws the compilers it installs from. Currently all the
urls seem to point to files hosted on <span><font color="#0000EE"><span style="white-space:pre-wrap"><u><a href="https://github.com/commercialhaskell/ghc/releases/" target="_blank">https://github.com/<wbr>commercialhaskell/ghc/<wbr>releases/</a></u></span></font><span style="white-space:pre-wrap"> and
from what I can tell there are far more ghc builds provided there
than directly produced by ghchq.</span></span></p>
<p><span style="white-space:pre-wrap">—gershom</span></p>
<p><br></p>
<div>
<blockquote type="cite" class="m_-2728895057694851186clean_bq" style="font-family:Helvetica,Arial;font-size:13px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none">
<div>
<div>
<div dir="ltr">
<div><span><br class="m_-2728895057694851186Apple-interchange-newline">
Cheers<br></span></div>
<div><span>Simon<br>
<br></span></div>
</div>
<div class="gmail_extra"><span><br></span>
<div class="gmail_quote"><span>On 12 May 2018 at 18:23, Gershom
B<span class="m_-2728895057694851186Apple-converted-space"> </span><span dir="ltr"><<a href="mailto:gershomb@gmail.com" target="_blank">gershomb@gmail.com</a>></span><span class="m_-2728895057694851186Apple-converted-space"> </span>wrote:<br></span>
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">
I believe this list/group is intended to help set GHC release
policies<br>
and plan when releases occur.<br>
<br>
I'd like to draw people's attention to<br>
<a href="https://ghc.haskell.org/trac/ghc/ticket/15068" rel="noreferrer" target="_blank">https://ghc.haskell.org/trac/g<wbr>hc/ticket/15068</a><br>

<br>
I think that the "nay" side has the right of it.<br>
<br>
1) To summarize my understanding: there is an issue with upstream
gnu<br>
binutils that are included in Ubuntu 18.04 and is triggered by
GHC<br>
_only_ when the -g flag for debug symbol output is passed. This bug
is<br>
triggered by GHCs going back through the 8.2 series. Note that
18.04<br>
was released _after_ ghc 8.4.2.<br>
<br>
There is a patch to GHC to work around this issue, that also<br>
incidentally improves another possible issue with debug
symbol<br>
emitting.<br>
<br>
Normally the -g flag is relatively rarely used, and by those who
would<br>
be able to work around such things. However, stack runs with -g
by<br>
default, unless users explicitly compile with symbol stripping
turned<br>
on.<br>
<br>
The proposed solution is an 8.4.3 release of GHC with _just_ this
patch.<br>
<br>
2) To summarize my objection to the plan -- I do not think
that<br>
spinning up a full new release to just patch one thing to work
around<br>
an upstream bug is a good precedent for GHC release management. I
can<br>
imagine exceptional circumstances, but in general the right
answer<br>
should happen downstream of a full GHC release. We already want a
more<br>
timely cadence. That cadence should suffice for these sorts of
things.<br>
<br>
In the meantime, the binary provided by hvr's ubuntu ppa
already<br>
incorporates this patch.<br>
<br>
One might say -- what about stack and platform users? My
suggestion<br>
would be that since stack provides its own binaries, patched
binaries<br>
can be provided by stack in this circumstance, just as the ubuntu
ppa<br>
includes the patch. Similarly we could respin a platform release
with<br>
the patch, just for linux, rather than across all three major<br>
platforms. (Or not -- I don't think that users that know about
-g<br>
would mind too much using the ppa for the patched release, and
I<br>
suspect those users are more likely ppa users already anyway, or
wise<br>
enough not to immediately upgrade to a new ubuntu system for
devwork<br>
without time for bugs to shake out).<br>
<br>
Anyway, I can't recall a precedent for a point-release over an
issue<br>
like this in the past, and I think if we start to do so too much
it<br>
will overstrain resources away from the important goals of
regularity<br>
and stability we do want in releases.<br>
<br>
I would like to know what others think.<br>
<br>
Cheers,<br>
Gershom<br>
______________________________<wbr>_________________<br>
Ghc-devops-group mailing list<br>
<a href="mailto:Ghc-devops-group@haskell.org" target="_blank">Ghc-devops-group@haskell.org</a><br>

<a href="https://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devops-group" rel="noreferrer" target="_blank">https://mail.haskell.org/cgi-b<wbr>in/mailman/listinfo/ghc-devops<wbr>-group</a><br>
</blockquote>
</div>
<br></div>
</div>
</div>
</blockquote>
<br class="m_-2728895057694851186Apple-interchange-newline"></div>
<br class="m_-2728895057694851186Apple-interchange-newline"></div>


</div></div></span></blockquote></div></div></div>
</blockquote></div><br></div>