Bug fixed compiler claim disputed | Drobe.co.uk archives

Bug fixed compiler claim disputed

By Chris Williams. Published: 9th May 2005, 00:11:53 | Permalink | Printable

Rewind that spin, say developers

A developer has challenged Castle's claim that all bugs reported against its C/C++ compiler package have been fixed.

In an announcement issued on Wednesday regarding the release of version 5.6 of the Castle C/C++ development tools, Castle stated that "this new release includes fixes for all bugs reported through Castle's C support web pages". Over the weekend, freeware programmer Tom Hughes alleged that this was "an interesting but unfortunately blatantly untrue claim, or at the very least an exceedingly misleading one."

Tom continued: "I can't actually find any bug reporting mechanism mentioned on the Castle web site, so on that basis the claim may, strictly speaking, be true as no bugs can be reported via a mechanism that does not appear to exist."

Tom says he has sent Castle details of seven bugs plus one enhancement request, of which two were fixed in last week's 5.6 release. Yesterday, Castle's John Ballance contacted Tom to ask for his bug reports again, adding that he'll check to see if there is a link to the bug reporting system from the compiler tools website.

Tom told us that he wasn't overly upset at Castle, remarking: "I just don't like marketing people bending reality in their press releases." He claims he has found two bugs in build tool Amu, two minor faults in the linker and bogus warnings in the compiler, including a new erroneous warning generated by the 5.6 compiler.

It is possible that Tom is not alone: one developer told us that reporting bugs to Castle can, in his opinion, sometimes feel like sending email into a black hole. Another programmer commented: "Pity the poor debuggers if not only are they expected to beat their heads against the bugs but respond to the public as well, though I suppose somebody ought to be doing it."

John explained to us that developers can report bugs via email, using an address supplied with the compiler package. He said: "There has been, over time, a number of bugs reported [via email]. The reports do not always get an immediate response, but the bugs do get captured. Some bugs are 'user' issues, but as with any complex piece of software, some holes do get exposed. No compiler is released unless it passes many qualifying tests, which includes building successfully the Iyonix ROM. However, different code trees and coders can and do occasionally expose code sequences that cause issues.

"What the press release states is that, (to the best of my own knowledge) all bugs so far reported (by that mechanism) are resolved."

Links

Castle website

Previous: New ROS powered kit hinted
Next: R-Comp sneaks out VRPC graphics speed boost

Discussion

Viewing threaded comments | View comments unthreaded, listed by date | Skip to the end

Well Thanks Chris! Castle bashing again.

However.. The statement is as far as I can determine Accurate in relation to the CC compiler.. Tom's bugs related in the main to amu (Make etc) and to the linker. I acknowledge that some bugs - or mis-features - relating to amu and to the linker are not yet fixed.

You will have noted that the 'tone' of the release was slanted heavily towards CC

John

jb on 9/5/05 11:24AM
[ Reply | Permalink | Report ]

"Castle bashing again." I'm not sure what you mean John; why "again". In fact, drobe has been accused of being biased towards Castle in the past. Chris is simply attempting to report the situation accurately. Naturally, you have a right to reply if you disagree.

The problem here is not so much the bugs in the compiler - some are expected in such a complex piece of software, but rather the opacity of Castle's bug reporting - both for the Compiler suite and for RISC OS 5 - noting bugs in your database (or whatever you use) might be great for the way you work, but it sadly doesn't give any great confidence to users, who don't even know if their bug is for real, for example. Perhaps Castle can consider ways in which this might be rectified. Ultimately, the time saved will be yours.

mrchocky on 9/5/05 11:41AM
[ Reply | Permalink | Report ]

Certainly, a central user-accessable bug tracking system for the whole OS and tool chain would save users a lot of time too - they could check to see if a bug they're experiencing has already been reported, and if there are any work-arounds, as well as it being handy for Castle, as mulitple people can provide input on the same bug, and it automatically all be kept in the same place.

nunfetishist on 9/5/05 2:13PM
[ Reply | Permalink | Report ]

I suppose what would be handy is a general knowledge base system that also includes bug reporting

sa110 on 9/5/05 7:43PM
[ Reply | Permalink | Report ]

Why not use Bugzilla?

JGZimmerle on 10/5/05 1:22AM
[ Reply | Permalink | Report ]

And is this something that just requires a bit of webspace and a little time to set up? Is it something that could be hosted here? Or is it a bit more of an undertaking? -- Spriteman

Spriteman on 10/5/05 8:54AM
[ Reply | Permalink | Report ]

Bugzilla is the kind of thing needed, yes. But it's generally considered that Bugzilla is hideously complicated, and that it actually makes it difficult for people to report bugs unless they're God.

nunfetishist on 10/5/05 11:42AM
[ Reply | Permalink | Report ]

See www.bugzilla.org for details.

On debian installation would be a simple #apt-get install bugzilla.

JGZimmerle on 10/5/05 11:49AM
[ Reply | Permalink | Report ]

Why would anyone call Bugzilla complicated? I think it is pretty intuitive. Anyone can try it on landfill.bugzilla.org.

JGZimmerle on 10/5/05 12:11PM
[ Reply | Permalink | Report ]

Are either the GCC or Castles C/C++ compilers capable of producing an Risc os version of Sibelius 3 . A stupid question I suppose but I just had to ask it.

jlavallin on 10/5/05 4:20PM
[ Reply | Permalink | Report ]

"Castle's", please. But your question misunderstands how these things work. GCC understands the full C++ specification (Castle C++ is old and does not), but the real issue is that any such version is likely to be tied to Windows, and need lots of work to be able to run on RISC OS, which is a much much larger problem than differences in compilers.

mrchocky on 10/5/05 4:28PM
[ Reply | Permalink | Report ]

I asked about Sibelius as I had been reading on their site as to the reason a Risc os version (Acorn) had not been produced was that the Acorn C++ compiler development had ceased in 1998. I assume from your reply that GCC's C++ compiler could be sufficient for Sibelius's use.

jlavallin on 10/5/05 5:23PM
[ Reply | Permalink | Report ]

jlavallin: If the compiler had been available in 1998, possibly. However, Sibelius 3 now is rather different from what it was seven years ago.

Hertzsprung on 10/5/05 5:41PM
[ Reply | Permalink | Report ]

I found that the most interesting reading in Castle's C changes logs concerns the CLib: the limit of 16 files open at once has finally been removed.

There are also mentions of "thread safe". I wonder if this work is at the same level as in GCC or if it is part of some wider work at OS level.

andretim on 10/5/05 7:14PM
[ Reply | Permalink | Report ]

Sibelius 7 being entirely in assembly language was one reason I was told why the Acorn version was discontinued. Unfortunately Sibelius 3 is very different. Apart from Firefox, a new RO Sibelius is something I would certainly love to see happen.

Sawadee on 10/5/05 9:29PM
[ Reply | Permalink | Report ]

Just to update people, John has now had details of my outstanding bugs from me and he has got back to me with his comments. One has been recategorised as an enhancement request and one (actually filed post 5.61 turns out to be incorrect). He has given me workarounds for some of the others as well as some indications as to how likely particular issues are to be fixed.

tom on 11/5/05 12:15AM
[ Reply | Permalink | Report ]

Sawadee: it is not going to happen. The RISC OS market is relatively tiny. The RISC OS music market is minute. They wouldn't even start to cover the costs of a new RISC OS Sibelius, never mind make money. Sadly, if you want to use new Sibelius software, you need a PC or a Mac.

To get the facts straight... a re-write of Sibelius was planned for Windows, MacOS and RISC OS. However, due to Acorn failing to support certain features in the compiler (I don't know what, nor do I understand programming well enough to discuss this issue) the Acorn version was dropped. Now, it is not commercially viable to produce a new Acorn version of Sibelius as anyone who wanted the new features moved platforms. After all, just about all professional users bought an Acorn to run Sibelius, not because they were Acorn fans. It's a safe bet that a lot only used the Acorn for composing and the PC/Mac for word processing and the like so would have no problem in switching to Sibelius for the PC or Mac.

arenaman on 11/5/05 1:17AM
[ Reply | Permalink | Report ]

Getting OT, but you would also have the problem of no MIDI output on the latest RISC OS hardware.

Although its a chicken and egg situation..noone is likely to produce MIDI hardware as there is no decent software to drive it. And noone is likely to produce some decent MIDI-capable software without any MIDI hardware.

Interestingly, the IYONIX has MIDI capability but only as far as pins on the motherboard which would need an interface card to be usable. However, it would probably be cheaper to use a mass produced PCI sound card to provide MIDI in/out that a custom interface card.

Cheers,

/Neil/

spellinn on 11/5/05 11:15AM
[ Reply | Permalink | Report ]

spellinn: 32-bit MIDI/MIDISupport drivers are basically finished. I've got a few bits to tidy up and then discuss with ESP how to release it. MIDI over USB is probably the easiest way forward for the new machines.

liquid on 11/5/05 11:37AM
[ Reply | Permalink | Report ]

liquid: Great to hear that, thanks! I agree that an USB MIDI interface is probably the best way, if someone could port an open source driver that would make things easier.

spellinn: I'd bet Stephen Borrill of MelIDI would consider recompiling it to be 26/32 bit neutral, as it is the only sequencer developed until fairly recently + it's the best sequencer for the platform.

People can tell eachother whatever they like, but the plain truth about Sibelius is that it became so powerful and popular, that its development was severely limited or hampered on the Acorn hardware and applications of that time. It simply outgrew our platform, just take a look at it and ask yourself if it had ever could have come this far on our platform, even with the Iyonix launched some years later. For an in-depth review, search for Sibelius3 on www.soundonsound.com - there's even some mention of its RISC OS roots

hEgelia on 11/5/05 4:30PM
[ Reply | Permalink | Report ]

Please login before posting a comment. Use the form on the right to do so or create a free account.

Search the archives

Today's featured article

Adventures with a Lego-cased A7K web server
Having previously built desktop and laptop cases of out Lego bricks, model building Peter Howkins has turned his attentions towards crafting a slim box to slid his A7000 into a rack, alongside other rackmount servers. Having pieced together the housing, Peter puts a legacy RISC OS machine through its paces as an internet-facing server.

11 comments, latest by jess on 3/12/08 2:07PM. Published: 21 Nov 2008

Random article

Teking a little longer than expected
Artex's Tek is delayed a while

6 comments, latest by jmcarey on 18/4/02 1:49AM. Published: 16 Apr 2002

Useful links

News and media:
Iconbar • MyRISCOS • ArcSite • RISCOScode • ANS • C.S.A.Announce • Archive • Qercus • RiscWorld • Drag'n'Drop • GAG-News

Top developers:
RISCOS Ltd • RISC OS Open • MW Software • R-Comp • Advantage Six • VirtualAcorn

Dealers:
CJE Micros • APDL • Castle • a4 • X-Ample • Liquid Silicon • Webmonster

Usergroups:
WROCC • RONE • NKACC • IRUG • SASAUG • ROUGOL • RONWUG • MUG • WAUG • GAG • RISCOS.be

Useful:
RISCOS.org.uk • RISCOS.org • RISCOS.info • Filebase • Chris Why's Acorn/RISC OS collection • NetSurf

Non-RISC OS:
The Register • The Inquirer • Apple Insider • BBC News • Sky News • Google News • xkcd • diodesign

© 1999-2009 The Drobe Team. Some rights reserved, click here for more information
Powered by MiniDrobeCMS, based on J4U | Statistics
"I was actually referring to drobe itself as the 'young upstart' rather than the age of the contributors"
Page generated in 0.0637 seconds.