We are running a survey to help us improve the experience for all of our members. If you see the survey appear, please take the time to tell us about your experience if you can.
Hi guys!
Whether there is somebody who tried the new version of uVision4 based on the Scintilla editor? Such *** I yet did not see! Where illumination of labels? Where put illumination of keywords? Why memory windows and assembly do not change a font? What for a hogwash with Russian fonts why are not displayed in a memory window? Words are not present...
Igor//
Dear Pony279 Pony279! It is strange, but I expected more rough reaction to my post. I always loved uVision, but the new version shook me! I'm assembler programmer and me am very important style of the text editor. To new meetings! //Igor
the editors the Keil people (who should stick to what they are good at (compilers) provide are all pitiful. I have always used an 'outside' editor, right now now CodeWright. Instead of complaining about the editor, my complaint is that they do not provide a method for linking a good editor (made by editor people) to uvision.
Erik
Hello Erik,
Thank You for the detail information about uVision's alternatives, but nothing to do with me. There are some unresolved questions:
1. Why quantity of cycles at performance of branch instructions in a simulator do not match a physical SiLabs CIP51 kernel? For example, JZ instruction will be executed always for an equal time? I would like to debug my assembly subroutines before I'll produce them. uVision lies here.
2. I'm a lover of uVision simulator, but where is a bit-wising display?
3. Why I cannot adjust a format of representation for each variable separately? For example, in Phyton environment this problem has resolve very good! In colour, besides!
4. I'm a Russian programmer, I would like to see a Russian fonts in my memory window to debug an LCD interface (I use XRAM-memory buffer to periodic update the LCD).
5. Where is a settings of period to source file backup? It's very important for every programmer!
6. Actually, so, there are any more questions to Keil team about how to do their product better. I think, You will force this process..
PS: Sorry for my bad English.
I do not use simulation. so 1-3 I'll leave alone. By the way, since you use SILabs why do you even simulate? I know no simulator for any processor that works 'perfect'. The JTAG debugging is always true.
Keil makes an excellent principal product, their principal product is not the editor. All the above is solved by using an editor made by editor people as opposed to limping editors made by compiler people. NOTE: I have the exact same complaint about IAR.
Erik: Can the debug interface on the Silabs chip count clocks? If it can't then it isn't a solution for someone who wants to count exact time consumption for a code block.
"Where is a settings of period to source file backup? It's very important for every programmer!"
Having an editor do automatic backup isn't a very important feature. It can even be a directly bad feature, compared to the alternative solution: an editor keeps a backup of the old file when saving.
When doing changes to code, I prefer to select my own save points when the code is in a stable state. With automatic backup, it can be saved when I have removed a source line from one location, but before I have pasted it into the new position. So I get a dangerous backup copy with a bug introduced.
Not only that - I want to do small changes and commit them so that it is easy to look at individual commits and identify if they introduced a behaviour change of the code or was just a quality-improvement of the code - maybe in preparation for introducing new functionality. With big commits, it isn't obvious which code changes that relates to which part of the check-in comments.
Another question here - what do you need that auto-backup for? Your computer regularly fails? My machine only reboots when M$ updates requires me to explicitly reboot. And the machine does not hang on its own. Last 12 months I have had uVision goof once, resulting in uVision dying unexpectedly. And that happend during a state change (start of a simulation) and not while editing code.
But yes - the editor component of uVision is lacking. Luckilly, I do most work with an older version of uVision before they switched editor component. The old one was also lacking but with less actual bugs.
the editors the Keil people (who should stick to what they are good at (compilers) provide are all pitiful
I don't think compilers is their strong point either. By the way, the ARM compiler is actually RealView (basically made by ARM.) They do have nice debuggers, rather accurate simulators covering lots of chips, and overall integration is good (and it is a lot of work.) What annoys me is how development tools vendors keep reinventing the same wheel - the editor. It ends up an utter failure every time. They should probably converge towards a common IDE (Eclipse?) It does seem like this is beginning to happen. But Eclipse has its own problems, I guess. Another possible solution would be for the vendors to use a proven high-quality editor component from a third party. As for me, I use an external editor and Makefile for building. I only have to run the native IDE for debugging. I can see the value of the native IDE for novice programmers, or for small projects. For large projects, it will only be a liability.
What is it about developers that get emotional on editors. Maybe I do something wrong, but I am effectively typing/editing code 10% of my development time. How much time can the perfect editor save me in reality? I want an editor to give me proper view and navigation of existing code. In the uVision editor I am mainly missing some Intellisense type feature. But that wouldn't let me enrage like the original poster.
I have yet to find the perfect editor!
98% of the time, I'll just use whatever's to hand.
Occasionally, when there is a big task to which one editor is particularly well suited, I'll switch editors for that task.
eg, CodeWright is the best I've found for column edit.
Erik: Can the debug interface on the Silabs chip count clocks? If it can't then it isn't a solution for someone who wants to count exact time consumption for a code block. set a timer to zero at the beginning set a breakpoint at beginning and end and you get the execution time including cache misses
"set a timer to zero at the beginning set a breakpoint at beginning and end and you get the execution time including cache misses"
Have you verified it? It very much depends on how the timers are handled when the processor core handles a breakpoint. To get usable values, the timer clock input must be gated by the breakpoint logic so the timers locks up at exact same time the program stops.
Many chips are totally unable to use breakpoints for measuring time. The best they can do is to have the code instrumented with actual processor instructions to read out the raw timer values before/after.
CodeWright or Coderight or Codewright. I now see why I so often mis-spell their name. My question. Is there a new release since Borland bought and killed Codewright?
Notepad++ has many features.
Bradford
Is there a new release since Borland bought and killed Codewright?
Not to my knowledge. 7.5 seems to be the end.
I have a suspicion that someone with a BIG say at Borland is an inferior editor maker. Borland first killed Brief, then CodeWright.
Some like to hide their own lack of abilities by attacking those of others.
Not so far as I know.
Last time I asked, there was no support and no development - yet they still thought it reasonable to charge full price if you wanted a copy!
"Notepad++ has many features"
Quantity != Quality
Althought I have recently been using Notepad++ and found it pretty good.
How much time can the perfect editor save me in reality?
nothing, if you do not care.
lots if you do.
e.g. if you 'steal' some of your code to use in a modified fashion elsewhere. a REAL editor is a fantastic timesaver getting formatting etc to be right after deleting and inserting the needed modifications.
also a REAL editor, has far better search abilities than a "compiler maker provided" one which can be a timesaver during debugging.
also, if you use some manufacturer supplied code (as a base), with a REAL editor it is a breeze to change the formatting to your standard, with a "compiler maker provided" one you almost may as well retype.
the $300 or so for CodeWright are recaptured in a month or two.
Absolutely.
Having the ability to move about files quickly and efficiently with a minimum amount of keypresses can be such a major timesaver. Also, using the decent one editor in this way means you don't have to keep hitting the slight differences that inevitably exist between different IDEs.
I started with Brief (when it was owned by underware, well before Borland got their mitts on it) and have now standardised on SlickEdit. I can now even go between different platforms with that same efficiency. Really nice.
I've just found that an updated version of Brief is still available at http://www.briefeditor.com/
Another possible solution would be for the vendors to use a proven high-quality editor component from a third party.
Which appears to be exactly what Keil did in a recent update to uVision: they switched to Scintilla, if memory serves. And look what it got them: even louder complaints than they used to receive for their own editor in recent times.
Lesson learned: there really is no way of getting this right to everyone's satisfaction.
I'm convinced that the Eclipse way is the right general idea. Given that there's almost certainly no such thing as the one, perfect IDE that everyone will accept without reservations, let's at least settle on the next best thing: one entirely usable, full-featured IDE core that everybody can use free of charge, and build their own extensions on top of. The world doesn't need more than one, maybe two of those.
Sure, Eclipse does have its own quirks and limitations, but at least it has only one set of those, which allows one to eventually find work-arounds for, or learn to accept them. But that's still way better than having dozens of IDEs each with its own set of different quirks and limitations, not to mention the truly inexcusable blunders each of those seemingly has to have at least one of.