Web app support still very broken

I am still very disappointed to see that 5.0.1 did not fix any of the major web app problems. Specifically:

1) The wrong web.xml is copied over to the exploded dir sometimes: if ModuleX depends on ModuleY, sometimes ModuleY's web.xml is copied. The only way to fix this is to open up the exploded dir, delete web.xml, and rebuild. This was claimed to be fixed before 5.0 was released.

2) With multiple web app modules structure with complex dependencies (though no circular dependencies) such as X->Y, Z->Y, and W->Y, without changing a single file the copy procedure can happen unnecessarily. This was a bug that I was told was resolved in 5.0.1. To be fair: it seems to be somewhat fixed (at least with my test case), but is still problematic.

3) This is a bug I haven't reported, but it is still important: when ModuleX depends on ModuleY and ModuleZ, collisions amongst the JSPs are not resolved using the order defined in ModuleX's Order/Export tab. Rather, they appear to be either random or alphabetical (ie: the order listed in the project).

Overall, I've found the quality around this feature is very bad. I took time out to restructure our project files to use this feature with the assumption that everything was fixed in 5.0.1. However, after a couple days of playing with it and more than half the engineers running in to these various problems, I can no longer depend on this feature.

We have moved to a custom Jetty launcher that looks directly in our sources. I hope that at some point JetBrains can do a review of the web module deployment code -- it needs it very much.

Patrick

PS: I don't mean to be a drag -- IDEA is still by far the only editor I will use... it is just too good. But I do feel that there are still some nasty bugs that crept through in this release:

- web module support (see above)
- the VM param "-cp" is not honored and breaks things when you run or debug applications. This has been broken in EVERY single release -- PLEASE write an automated test for this. Because this doesn't work, we can't even launch Resin the way we used to, which provides fewer alternatives in light of the broken web module support.
- when specifying an alternative JRE, the same classes are not put in the classpath (namely dependent module's outputs).
- javascript and css support, while in the right direction, are effectively worthless until you can solve the issue that they are typically applied dynamically. I know, it's a hard problem, but it is very important. Even simple solutions, such as specifying URL patterns to map to JS/CSS resources, would be better than what we have now.

16 comments
Comment actions Permalink

Here's the bug about redeploying unnecessarily:

http://www.jetbrains.net/jira/browse/IDEADEV-2263

0
Comment actions Permalink

Hello,

We also use dependant Webapp modules heavily and the recent version 5.0 of
IDEA is very buggy w.r.to to handling these dependencies. We are seeing similar
problems. Please fix them ASAP - this is preventing us from upgrading to
the latest version (and use other great features).

Regards,
--Venkat.

PL> I am still very disappointed to see that 5.0.1 did not fix any of
PL> the major web app problems. Specifically:
PL>
PL> 1) The wrong web.xml is copied over to the exploded dir sometimes:
PL> if ModuleX depends on ModuleY, sometimes ModuleY's web.xml is
PL> copied. The only way to fix this is to open up the exploded dir,
PL> delete web.xml, and rebuild. This was claimed to be fixed before 5.0
PL> was released.
PL>
PL> 2) With multiple web app modules structure with complex dependencies
PL> (though no circular dependencies) such as X->Y, Z->Y, and W->Y,
PL> without changing a single file the copy procedure can happen
PL> unnecessarily. This was a bug that I was told was resolved in 5.0.1.
PL> To be fair: it seems to be somewhat fixed (at least with my test
PL> case), but is still problematic.
PL>
PL> 3) This is a bug I haven't reported, but it is still important: when
PL> ModuleX depends on ModuleY and ModuleZ, collisions amongst the JSPs
PL> are not resolved using the order defined in ModuleX's Order/Export
PL> tab. Rather, they appear to be either random or alphabetical (ie:
PL> the order listed in the project).
PL>
PL> Overall, I've found the quality around this feature is very bad. I
PL> took time out to restructure our project files to use this feature
PL> with the assumption that everything was fixed in 5.0.1. However,
PL> after a couple days of playing with it and more than half the
PL> engineers running in to these various problems, I can no longer
PL> depend on this feature.
PL>
PL> We have moved to a custom Jetty launcher that looks directly in our
PL> sources. I hope that at some point JetBrains can do a review of the
PL> web module deployment code -- it needs it very much.
PL>
PL> Patrick
PL>
PL> PS: I don't mean to be a drag -- IDEA is still by far the only
PL> editor I will use... it is just too good. But I do feel that there
PL> are still some nasty bugs that crept through in this release:
PL>
PL> - web module support (see above)
PL> - the VM param "-cp" is not honored and breaks things when you run
PL> or debug applications. This has been broken in EVERY single release
PL> -- PLEASE write an automated test for this. Because this doesn't
PL> work, we can't even launch Resin the way we used to, which provides
PL> fewer alternatives in light of the broken web module support.
PL> - when specifying an alternative JRE, the same classes are not put
PL> in the classpath (namely dependent module's outputs).
PL> - javascript and css support, while in the right direction, are
PL> effectively worthless until you can solve the issue that they are
PL> typically applied dynamically. I know, it's a hard problem, but it
PL> is very important. Even simple solutions, such as specifying URL
PL> patterns to map to JS/CSS resources, would be better than what we
PL> have now.


0
Comment actions Permalink

At this point I've given up on IDEA's exploded directory support and the various plugins (JBoss, Resin, Tomcat, etc). I have no desire to fill out anymore bug reports, as it has taken considerable time already and it is clear the results weren't worth my effort.

I know JetBrains is working hard to fix these, and I know they've gone through some serious growth in the last two years. But it is frustrating to see the same bug happen on every single non-bug-fix release since 3.0 (classpath bug in VM params, as mentioned above).

With HotSwap crippled/slow, the recurring VM params bug, and broken web module support, I'm so far not thrilled about this 5.0 release. Add on the fact that CSS and JSP support is just a "tease" by the very nature of how HTML is typically dynamically generated, and I'm just frustrated.

On top of that, JSP editing support is vastly slower, and my OS X coworkers have stability problems hourly. Please, JetBrains, do what you do best: make a kick ass product again!

0
Comment actions Permalink

Regarding #2 mentioned by Patrick these files which are copied are *.properties and *.xml and *.xsl in my case. I guess not *.class files are still copied on every build.

Please fix these issues.

Thanks,
--Venkat.

0
Comment actions Permalink

I've got to agree with you here - I've resorted to writing an Ant script to carry out my deployment, as my IDEA constantly loses deployment settings and/or doesn't copy across all libraries. I don't mind fixing bugs in my own code, but when a bug is being caused because a library hasn't been copied yet again, it gets very frustrating.

The deployment settings for webapps with a lot of dependent modules are now being lost virtually every day - I can't reproduce the bug reliably however. Having to reset which of the ~120 libraries are supposed to be being copied just got too irritating. This bug was in version 4.5 of IDEA as well - I was really hoping that webapp deployment would improve in this version and not get worse.

0
Comment actions Permalink

Venkat, what is the SCR number of the issues you are talking about ?

--
regards,
--
Alexey Kudravtsev
Software Developer
JetBrains, Inc, http://www.jetbrains.com
"Develop with pleasure!"


"Venkat Sonnathi" <no_mail@jetbrains.com> wrote in message
news:31822730.1125469584691.JavaMail.javamailuser@localhost...

Regarding #2 mentioned by Patrick these files which are copied are
*.properties and *.xml and *.xsl in my case. I guess not *.class files are
still copied on every build.

>

Please fix these issues.

>

Thanks,
--Venkat.



0
Comment actions Permalink

Yup totally agree. I've started using netbeans more heavily because it
doesn't piss me off every time I want to compile a web app, or hotswap a
class, or just type without having to wait for the editor to catch up
with me, or when I hit return wait a couple of seconds for idea to make
up its mind on whether it's ready to act on the return or if it needs to
go to lunch first.

The only reason I'm still in idea is because the refactoring tools are
so much better, but I hear there are other free tools out there for
netbeans and I'm considering it seriously this time.

There is NO reason, and I've said it so much JetBrains probably hates me
for it, that you can edit so incredibly fast in netbeans on OSX, and yet
in idea it's slow as molasses. If it's because of some parsing idea is
doing and blablabla, fine then give us a button to turn that off so we
can get full speed ahead. However, I know that can't be it, because
netbeans does quite a bit of parsing now, and even has the little
green/red square at the top right of the editor to indicate the status
of a file... just like idea.

If you look at Jira, it's been a long while since I've filed
bugs/feature requests. The only thing I've submitted has to do with
exceptions. I don't plan on filing any more issues until I see some
serious speed improvements on OSX's editing side.

Best.
R

0
Comment actions Permalink

If it's because of some parsing idea is
doing and blablabla, fine then give us a button to turn that off so we
can get full speed ahead.

I thought Hector was exactly for this purpose there :).
Have you tried it and it's still slow?

Ahmed.

0
Comment actions Permalink

Simon -- we have the exact same problem!!! I forgot about that one too, but in the last week we've had that problem hit one of our developers probably more than two times a day. Very annoying!

0
Comment actions Permalink

In article <df49c0$eut$1@is.intellij.net>,
Ahmed Mohombe <amohombe@yahoo.com> wrote:

If it's because of some parsing idea is
doing and blablabla, fine then give us a button to turn that off so we
can get full speed ahead.

I thought Hector was exactly for this purpose there :).
Have you tried it and it's still slow?


Yup does NOTHING to improve speed. 0.

R

0
Comment actions Permalink

>>>If it's because of some parsing idea is
>>>doing and blablabla, fine then give us a button to turn that off so we
>>>can get full speed ahead.
>>I thought Hector was exactly for this purpose there :).
>>Have you tried it and it's still slow?

Yup does NOTHING to improve speed. 0.

Well, than you should really send to Jetbrains some more profiler
snapshots (in different contexts).

If you have a 'test project' where this is reproducible, than send it
too, cause maybe Jetbrains is using too simple 'test projects' that do
not reflect the nowadays web projects situation.

However, I think many of the opensource projects are good enough to
evidence that "Develop with Pleasure" can't be really applied to web
projects :(.

Ahmed.

0
Comment actions Permalink

If you have a 'test project' where this is reproducible, than send it
too, cause maybe Jetbrains is using too simple 'test projects' that do
not reflect the nowadays web projects situation.


I've sent more than I can count during EAP. I'm done sending in more
files which will then be looked at and then I'll be told that this is
all they can do, or that 1 second is OK for a return, or that the typing
speed on their end is not slow... etc...

I've spent more time debugging IDEA than I care to do it sometimes for
my own projects. It's silly.

R

0
Comment actions Permalink

Hello Alexey,

Here it is http://www.jetbrains.net/jira/browse/IDEADEV-2263

and step 6 (in the list of steps mentioned) talks about copying the files
over even though none of them changed. It certainly has improved but still
some copying goes on if you press Ctrl F9 and after successful build, press
Ctrl F9 again you will see some files copying over which are *.properties,
*.xml and *.xsl in my case.

Thanks for your time and please make the depedant webapp feature robust and
a pleasure to use like the rest of the other features.

Regards,
--Venkat.

AK> Venkat, what is the SCR number of the issues you are talking about ?
AK>
AK> "Venkat Sonnathi" <no_mail@jetbrains.com> wrote in message
AK> news:31822730.1125469584691.JavaMail.javamailuser@localhost...
AK>
>> Regarding #2 mentioned by Patrick these files which are copied are
>> *.properties and *.xml and *.xsl in my case. I guess not *.class
>> files are still copied on every build.
>>
>> Please fix these issues.
>>
>> Thanks,
>> --Venkat.


0
Comment actions Permalink

In article <robert-4F9F65.08500231082005@mail.intellij.net>,
"Robert S. Sfeir" <robert@NOSPAMCodepuccino.dom> wrote:

Yup totally agree. I've started using netbeans more heavily because it
doesn't piss me off every time I want to compile a web app.


By the way - if you want to use the JFluid profiler, file a bug at Apple
asking for a vm rev to 1.5.0_04. Also note specifically that JFluid is
something you want to have work.

Apparently, this is the first version of 1.5.0 that will support JFluid
at the VM level.

Scott

--
Scott Ellsworth
scott@alodar.nospam.com
Java and database consulting for the life sciences

0
Comment actions Permalink

In article <robert-4F9F65.08500231082005@mail.intellij.net>,
"Robert S. Sfeir" <robert@NOSPAMCodepuccino.dom> wrote:

There is NO reason, and I've said it so much JetBrains probably hates me
for it, that you can edit so incredibly fast in netbeans on OSX, and yet
in idea it's slow as molasses. If it's because of some parsing idea is
doing and blablabla, fine then give us a button to turn that off so we
can get full speed ahead. However, I know that can't be it, because
netbeans does quite a bit of parsing now, and even has the little
green/red square at the top right of the editor to indicate the status
of a file... just like idea.


If JetBrains can pin down any performance problems as being Apple's
fault, I know that Apple wants to hear about them really soon. Given
how speedy NetBeans is, I have my guess, but JetBrains can find out for
sure.

I talked with the Java team at Apple last week. They were very
interested in any case where their VM is slower than a corresponding
Windows machine. If JetBrains can produce profiles that demonstrate
Apple problems, and then file appropriate bugs with Apple, they will get
attention.

The ideal bug is a small JUnit test case, as this can be put into their
nightly test regime, but they said they will take what they can get.
Even a sample/Shark profile of a running app doing something that takes
too long is a help.

Remember - by WWDC 2006, they will have a machine that can boot Windows
as well as the MacOS. Their VM will be compared with the Windows VM on
exactly the same hardware. If they look bad, it will not be a happy day
in Mac land. Any problems they can find and fix now are ones they will
not have to fight later.

According to one person, Java performance is now on the radar at the
highest levels - the top people at Apple do not want the Intel
transition to flop, and having Windows be better at anything major will
really hurt adoption of the new machines. The Apple Java team wants to
know what needs to be fixed, and they want to know soon.

There is a major OS release due near the end of 2006, so some things may
get pushed back, but however it goes, they want a good VM, and that
requires samples of the not so good parts.

So - JetBrains - if the slowdowns are Apple's fault, please tell them.
Also, as a courtesy, tell us, so we can file similar bugs to let them
know that we are paying attention.

Scott

--
Scott Ellsworth
scott@alodar.nospam.com
Java and database consulting for the life sciences

0
Comment actions Permalink

On Wed, 31 Aug 2005 18:37:51 -0400, Scott Ellsworth <scott@alodar.com>
wrote:

So - JetBrains - if the slowdowns are Apple's fault, please tell them.
Also, as a courtesy, tell us, so we can file similar bugs to let them
know that we are paying attention.


Now THAT is something I would be more than willing to help with. I'm sure
Max is reading the forums again, and Max knows well of the issues I've
brought up many many times, his ear has rotted because of my complaints.

I can list a few things which are slower, perhaps Jetbrains can put in the
company to blame next to it:

1- Typing in jsp is slow, anywhere not just when the line is the longest
2- Typing is very slow when the line is the longest, in any editor window
3- Hitting return is slow in jsps. Pressing return down in fast sequence
takes a long time to finish the task.
4- Scrolling in a document is slow, scroll up and hold the arrow key up,
taking finger off, watch screen scroll for another few seconds.
5- Opening a jsp file is not as zippy as on windows. Takes a few seconds
on OSX, windows feels almost instantaneous.
6- Showing/Hiding panes like project, ant etc... redraw is jittery. On
windows it's quick and crisp. It's even worse when you start a web app,
and you need to compile, it feels like molases when the compile windows
comes up, then hides, then the hotswap window comes up, then hides then
you get back to the debug window. The swapping between window feels like
for ever between swaps.
7- Some of the IDE preferences takes a few seconds to show up, not just
appearance which has to build a list of fonts.
8- When changing the font of the IDE, the main menu duplicates over and
over for every font change, you end up seeing nothing but .... at the top
because the menu duplicates so many times. Clicking OK to set the font
change returns the menu to normal.
9- When setting code style, there are some dropdowns you can choose to
change a style, like where you set the wrapping, you have to click that
dropdown 2-3 times before you get the choices to stick.

I don't exactly have a crappy box. I'm on a 1.5Ghz PB 17" with 128MB of
VRAM and 1.5 Gig of RAM, and the faster 80 gig drive. It's not a dual G5,
but still I can say that a majority of Mac users have this type of setup
v/s the dual box.

R

0

Please sign in to leave a comment.